When developing a wellbore stability analysis project, it is important to have as much information about the area to be drilled as possible due to the importance of data accuracy to construct the mud weight window. Usually, in order to avoid drilling events, data of adjacent wellbores are used to deterministically estimate the fracture and collapse gradients and stablish a mud weight range that assures the wellbore stability along the desired trajectory. This paper presents a wellbore stability analysis based on a probabilistic method -First Order Second Moment (FOSM) -that considers a statistic analysis to characterize the uncertainties associated to the failure pressures calculated. First, a deterministic method is used to calculate the fracture and collapse gradients. The input parameters are assumed to have a standard deviation of 10% of the mean value. Then, the results of the first step are used as input data to the probabilistic method that will provide scenarios for different probabilities of success (P10, P50 and P90). The probabilistic method proposed is based on a statistical analysis to define the standard deviation associated to the gradients' calculus. This paper will approach two different ways to define this standard deviation: normal variation and random variation. Once the analysis of the results is made, it is possible to conclude for the probabilistic method which scenario presents a more conservative behavior, the standard deviation of the results and it will be possible to compare the two statistical analysis described for each of the failure model.

RESUMO: Na indústria do petróleo, é fundamental uma boa estimativa da resistência da rocha para uma análise adequada de estabilidade. Entretanto, a obtenção de testemunhos é uma operação dispendiosa. Para extraí-los é necessário interromper a perfuração, demandando tempo de equipe e aluguel de sonda. Sendo assim, é preciso adotar uma forma indireta para prever a resistência das rochas, por meio dos dados obtidos dos perfis, aplicando correlações adequadas. No caso das rochas carbonáticas, essa avaliação torna-se mais complexa, pois envolve o estudo de estruturas heterogêneas. Esse trabalho apresenta uma estimativa de resistência para carbonatos análagos aos do pré-sal brasileiro, com base em conceitos de classificação de rochas carbonáticas, análise crítica de correlações existentes na literatura e comparação com resultados de testes de laboratório. A partir da avaliação da correlação de Prasad e do estudo granulométrico elaborado para rochas carbonáticas diversas, é proposta uma metodologia para estimar os parâmetros de resistência máximo e mínimo de cada litotipo analisado. Dada a escassez de ensaios de laboratório para o contexto de poços de óleo e gás, a definição de uma faixa de valores permite uma abordagem probabilística que auxiliará na previsão da resistência da rocha.

No abstract

No abstract

Wellbore stability studies are essential to guarantee the success during drilling oil and gas wells. This case study shows a Post Morten analysis of a horizontal well successfully drilled along a depleted reservoir, but with a narrow mud weight window, after a lost circulation during the cementation of the previous casing. The main goal is to show that even having a well-established general methodology for wellbore stability analysis, specific challenges may require different solutions to calibrate the mud weight window with field evidences. Challenges faced during drilling this wellbore were: (i) the wellbore was drilled after a lost circulation during cementing the previous phase. (ii) the existence of a narrow mud weight window, conditioned by unconsolidated sands with low resistance parameters, in the reservoir; (iii) the well path implies a critical combination of stresses around wellbore wall; (iv) the presence of a depleted reservoir. The in-situ stress prediction of horizontal stresses involved different criteria and showed a strong difference between the overburden and the reservoir. Minifrac tests, carried out in many wells along the field, present a considerable variation along the reservoir, suggesting some in-situ stress variation throughout the field. The analysis imposed a series of iterations by a carefully guided calibration of the input parameters and failure criteria. The analysis allowed calibrating and explaining field evidences, including the losses during cementing.